Skip to content
notes · measurement

Your thin pages drag down your good ones

Structure and markup refreshed for current answer engines; the original analysis is preserved.

In late February Google launched the update the industry first called Farmer and now calls Panda — a change that hit content farms and thin, low-value pages hard, touching nearly 12% of queries. The part that matters most is how it judges: not page by page, but site by site. Panda scores quality across your whole domain, so a pile of weak, shallow, or duplicated pages can pull down the strong ones sitting beside them. The lesson is a measurement lesson. Your quality is an aggregate, not a highlight reel, so audit the whole inventory, find what is dragging, and improve it, merge it, or remove it.

the short answer

In late February Google launched Panda (first called Farmer), hitting content farms and thin pages and touching ~12% of queries. The new thing: it scores quality across the whole site, not page by page — entire domains fell. So thin content is harmful, not neutral: weak pages drag down the strong ones and lower your aggregate. Audit the whole inventory, then improve, consolidate, or remove what drags — don’t pile more on top.

key takeaways

  • In late February Google launched the update first called Farmer, now Panda — announced by Singhal and Cutts — hitting content farms and thin, low-value pages and touching close to 12% of queries.
  • The genuinely new thing: Panda scores quality at the level of the whole site, not just the page. Entire domains fell, not only the offending pages — a high proportion of weak pages pulls down the rest.
  • So thin content is harmful, not neutral. A pile of shallow, duplicated, or padded pages actively drags down the strong ones beside them, lowering the aggregate Google reads. An extra thin page is a small weight, not a free extra.
  • That inverts the content-farm instinct: publishing more shallow pages to target more terms is now a liability. A big site of strong pages is fine; a big site padded with thin ones pays a tax.
  • What to do: audit the whole inventory, measure the strong-to-weak balance, then improve, consolidate, or remove what drags. Don’t pile more on top. Recovery is gradual — quality is reassessed over time, not overnight.

the average is what gets measured

two sites, same strong pages — different averages Site A — padded with thin pages low average → whole site ranks down Site B — thin pages pruned improved / merged / removed high average → strong pages rise Panda judges the site, not the page. Thin content is a weight — lift the average.

The strong pages are identical in both sites. What differs is the company they keep — and under a site-wide measure, the company a page keeps is part of how it is judged.

The idea, in four parts

What happened in late February; why quality is judged across the whole site; why thin content is a weight rather than neutral; and what to do about it. Open each part.

01 What happened in late February

In late February Google rolled out a major change to how it ranks, announced on the official blog by Amit Singhal and Matt Cutts. In Google’s own framing, the update lowers the rankings of low-quality sites — those that add little for users, copy their content from elsewhere, or are simply not very useful — while raising sites that do original, substantial work: research, in-depth reports, thoughtful analysis. The industry named it first, calling it the Farmer update, because its most conspicuous victims were content farms — operations that paid writers a pittance to produce thousands of shallow articles targeting every search term imaginable. Google has since let slip that internally it goes by Panda, after the engineer behind the central breakthrough, and that is the name taking hold. The scale was striking: it noticeably affected close to twelve percent of queries, an enormous figure for one change, and the consequences were stark. Prominent content farms lost much of their visibility almost at once, and established sites known for reliable information climbed into the space they vacated. Some context helps explain the timing: Google’s faster indexing system, rolled out the year before, had made it easier for the swelling volume of shallow content-farm pages to enter the index and rank, sharpening a quality problem that critics had been complaining about loudly. Panda is the response. And as Matt Cutts has put it, the sites it targets are the ones asking what is the bare minimum I can do that is not quite spam — content that slipped past the older filters built to catch obvious junk. Rather than hunting rule-breakers, Panda assesses quality itself.

02 Quality is judged across the whole site

The genuinely new idea in Panda, and the one worth understanding well, is that it scores quality at the level of the whole site rather than the single page. Earlier signals largely weighed a page on its own; Panda assigns a site-wide quality assessment, and a domain carrying a high proportion of weak, shallow, or duplicated pages can see its rankings drop across the board. Entire sites took the hit — not merely the offending pages, but the good ones sitting alongside them. The reasoning holds up once you see it. If most of what a site publishes is thin, that says something about the site as a whole, and Google’s read on any individual page is coloured by the company that page keeps. This is exactly why content farms were so exposed: their entire model was volume, and a large volume of shallow pages is precisely what a site-wide quality measure is built to punish. It is also why there were some unintended casualties — sites with genuinely good content but a scattering of weak pages got swept up, because what moved them was the aggregate, not their best work. For anyone running a real site, the implication is uncomfortable and important: you cannot quietly wall off your weak content and trust your strong pages to carry the domain. The weak pages are part of the average, and the average is part of what Google now measures. The unit of judgement has shifted from the page to the site, and that changes what you have to pay attention to.

03 Thin content isn’t neutral — it’s a weight

Follow the site-wide logic to its conclusion and you reach the part most people find hardest to accept: thin content is not harmless, it is harmful. The old intuition treated an extra shallow page as cost-free — at worst it does nothing, and it might rank for something, so why not publish it. Panda turns that on its head. Because quality is assessed across the whole site, a mass of shallow, duplicated, or low-value pages actively drags on the pages you are proud of, dragging down the aggregate Google reads off your domain. A thin page is no longer a harmless extra sitting off to the side; it is a small weight pressing on everything else you have published. This is what made the early Panda recoveries so counter-intuitive. For many sites, the fix was not to write more but to cut or seriously improve the weak content already there, because removing the dead weight let the strong pages rise. It also means the instinct that powered the content-farm years — publish more, chase more terms, pile up more pages — has flipped from asset to liability wherever those pages are shallow. To be clear, this is not an argument against large sites. A big site full of consistently strong pages is perfectly healthy; what Panda penalises is a big site padded out with thin ones, carrying a tax it never needed to take on. Seeing thin content as a real cost rather than a free option — and weighing that cost honestly against any traffic it brings — is the measurement discipline the AC Group has worked by for {years} years.

04 What to do about it

Treat your content as an inventory to be measured and managed, not a pile to keep adding to. Begin by looking at everything you publish, not just your best work — a site-wide measure means your worst pages count toward the verdict, so you have to actually see them. Go through the inventory and sort it honestly: which pages are substantial and genuinely useful, which are thin, duplicated, or padded, and which fall somewhere between. For each weak page, you have three honest options, and the page decides which fits. Improve it, when the topic deserves a real page and you can make it one. Consolidate it, when it is one of several near-duplicates that should collapse into a single strong page. Or remove it — delete it, or keep it out of the index — when it adds nothing and never will, because a page that does nothing but dilute your average is better gone than kept. Resist the reflex to publish more on top of the problem; with a site-wide quality measure, piling on thin pages drives the aggregate down, not up. And measure as you work: track the share of your site that is strong against the share that is weak, and watch the ratio improve as you prune and lift. One honest caveat about pace, drawn from how Panda operates — quality is reassessed over time rather than instantly, so the benefit of cleaning up arrives gradually, not the morning after. Managing your content by the numbers, as an inventory whose average you are responsible for, is the measurement discipline the AC Group has worked by for {years} years.

Why this is an inventory point, not a publishing one

The instinct a change like Panda provokes is to ask what to publish next, and that is the wrong question first. Before anything new, Panda is a statement about what you already have: your site is being judged as a body of work, and the weak parts of that body count. So the prior question is not what to add but what to look at — the whole inventory, including the pages you would rather not think about, because those are the ones moving your average.

Framed that way, Panda is less a penalty than a prompt to manage what you own. Most sites have accumulated pages that no longer earn their place — thin, duplicated, outdated, padded — and have never measured what those pages cost the rest. Taking the inventory seriously, sorting strong from weak and acting on the difference, is the unglamorous work that quality at the site level rewards. Doing that honestly, by the numbers, is the measurement judgement the AC Group has brought to clients for 27 years.

What to do with this

Audit everything, not just your best work. Sort the inventory honestly into strong, weak, and in-between, and for each weak page take one of three actions: improve it if the topic deserves a real page, consolidate it if it is one of several near-duplicates that should be a single strong page, or remove it — delete or de-index — if it adds nothing and never will. Resist publishing more on top of the problem; under a site-wide measure, more thin pages lower the average.

Measure as you go: track the share of your site that is strong against the share that is weak, and watch it improve as you prune and lift. And keep your expectations honest about pace — quality is reassessed over time, so the gains from cleaning up arrive gradually rather than overnight. Managing your content as an inventory whose average you are responsible for, rather than a pile to add to, is the measurement discipline the AC Group has worked by for 27 years.

Panda, plainly: quick answers

What is the Panda update?

In late February Google rolled out a major ranking change — Amit Singhal and Matt Cutts announced it on the official blog — that it said would reduce the rankings of low-quality sites, the ones that add little value, copy content from elsewhere, or simply are not very useful, while lifting sites with original, substantial work like research, in-depth reports, and genuine analysis. The industry first named it the Farmer update, because its most visible casualties were content farms: sites that paid writers very little to churn out thousands of shallow articles aimed at every conceivable search term. Google has since let on that internally it was called Panda, after the engineer behind the core breakthrough, and that name is sticking. The scale was large — it noticeably affected close to twelve percent of queries, which is enormous for a single change — and the effects were dramatic: well-known content farms lost much of their visibility, while established sites known for solid information rose to take their place. Matt Cutts has described the kind of site Panda targets as one asking what is the bare minimum I can do that is not quite spam, slipping past the older filters that caught obvious junk. Panda is the answer to that gap: rather than catching rule-breaking, it assesses quality, modelled on how human raters would judge whether a site is trustworthy and worth reading. The headline is that Google got materially better at telling shallow content from substantial content, and acted on the difference at scale.

Why does Panda judge my whole site, not just a page?

Because that is the genuinely new thing about it: Panda scores quality at the level of the whole site, not only the individual page. Earlier ranking signals largely judged a page on its own merits, but Panda assigns a site-wide quality assessment, and a domain carrying a high proportion of weak, shallow, or duplicated pages can see its rankings fall across the board — entire sites took the hit, not just the offending pages. The logic is reasonable once you see it. If most of what a site publishes is thin, that is evidence about the site as a whole, and Google’s judgement of any one page is coloured by the company it keeps. This is why content farms were so exposed: their model was volume, and volume of shallow pages is precisely what a site-wide quality measure punishes. It is also why some sites with genuinely good content but a scattering of weak pages were caught in the net — the aggregate, not the best page, is what moved them. For anyone running a real site, the implication is uncomfortable but clear: you cannot quarantine your weak content and assume your strong pages will carry you. The weak pages are part of the average, and the average is part of what Google measures. Understanding that quality is assessed in aggregate rather than page by page changes what you have to look at, and looking at the whole picture rather than the highlights is the kind of measurement discipline the AC Group has practised for 27 years.

Is thin content really harmful, or just unhelpful?

With a site-wide quality measure in play, it is harmful, not merely neutral — and that is the shift people find hardest to absorb. The old intuition was that an extra thin page could only help or, at worst, do nothing: more pages, more chances to rank, no downside. Panda inverts that. Because quality is assessed across the whole site, a large body of shallow, duplicated, or low-value pages actively drags on the pages you are proud of, lowering the aggregate that Google reads. A thin page is no longer a harmless extra; it is a small weight on everything else. This is what made Panda recoveries so counter-intuitive at first: the fix for some sites was not to add more content but to remove or substantially improve the weak content already there, because cutting the dead weight lifted the rest. It also means the instinct that served the content-farm era — publish more, target more terms, accumulate more pages — is now a liability rather than an asset wherever those pages are shallow. None of this argues against having a large site; it argues against having a large amount of weak content on it. A big site of consistently strong pages is fine; a big site padded with thin ones is carrying a tax it does not need to pay. Recognising thin content as a cost rather than a free option, and weighing it honestly, is the measurement discipline the AC Group has worked by for 27 years.

So what should I actually do?

Audit the whole inventory, measure the balance of strong to weak, and then improve, consolidate, or remove what is dragging. Start by looking at everything you publish, not just your best work — the point of a site-wide measure is that your worst pages count, so you have to see them. Go through the inventory and sort it honestly: which pages are genuinely substantial and useful, which are thin, duplicated, or padded, and which sit in between. For the weak pages, you have three honest options, and the right one depends on the page. Improve it, if the topic deserves a real page and you can make it one. Consolidate it, if it overlaps several near-duplicates that should be a single strong page. Or remove it — delete or keep it out of the index — if it adds nothing and never will, because a page that only dilutes your average is better gone. Resist the reflex to simply publish more on top of the problem; with a site-wide quality measure, adding thin pages makes the aggregate worse, not better. And measure as you go: track what share of your site is strong versus weak, and watch that ratio improve as you prune and lift. One caution drawn from how Panda works — recovery is not instant; quality is reassessed over time, so the gains from cleaning up show up gradually rather than overnight. Treating your content as an inventory to be measured and managed, not a pile to be added to, is the measurement discipline the AC Group has worked by for 27 years.

A note on sources and timing

This is written in early 2011, just after the update. The description — that Google launched a major ranking change in late February, announced by Amit Singhal and Matt Cutts, aimed at low-quality sites and content farms while rewarding original, substantial work; that the industry first dubbed it Farmer and Google’s internal name Panda is taking hold; that it affected close to twelve percent of queries; that its defining feature is a site-wide quality assessment so that whole domains, not just individual pages, rose or fell; and that Google’s faster indexing the year before had sharpened the content-farm problem — follows Google’s statements and contemporaneous reporting. The reading offered here — that quality is measured in aggregate, that thin content is therefore a weight rather than a neutral extra, and that the response is to audit the inventory and improve, consolidate, or remove what drags — is our interpretation, grounded in that record. The durable point outlasts this update: your thin pages drag down your good ones, so measure and manage the whole. That is the measurement discipline the AC Group has worked by for 27 years.

Do you know what your weak pages are costing you?

Our free audit measures your content as an inventory — the balance of strong to thin across your whole site — and shows where weak, duplicated, or padded pages are dragging down the ones that should rank. In English and Spanish, in 48 hours, with no sales call.